Improving Mixed-Criticality System Consistency and Behavior on Multiprocessor Platforms by Means of Multi-Moded Approaches
Recent research in the domain of real-time scheduling theory has tackled the problem of scheduling mixed-criticality systems upon uniprocessor or multiprocessor platforms, with the main objective being to respect the timeliness of the most critical tasks, at the expense of the requirements of the less critical ones. In particular, the less critical tasks are carelessly discarded when the computation demand of (some of) the high critical tasks increases. This might nevertheless result in system failure, as these less critical tasks could be accessing data, the consistency of which should be preserved. In this paper, we address this problem and propose a method to cautiously handle task suspension. Furthermore, it is usually assumed that the less critical tasks will never be re-enabled once discarded. In this paper, we also address this concern by proposing an approach to re-enable the less critical tasks, without jeopardizing the timeliness of the high critical ones. The suggested approaches apply to systems having two or more criticality levels.